The Characteristics of Attention Flow of MOOC Students and Their Prediction of Academic Persistence

Through analysis of a typical MOOC click-stream data, this paper explores the attention flow characteristics of MOOC students, compares and analyzes the attention flow characteristics of students with different performances, and proposes a real-time prediction model of academic persistence based on marked attention flow temporal data by using Recursive Neural Network combined with LSMT algorithm. After testing, the predictive model achieved the accuracy of 0.84. This paper provides an insight into MOOC's curriculum design and real-time learning monitoring and early warning based on click-stream data.